CVE-2026-6539: Notepad++ format string bug enables

CISO Take

This is a local format string injection (CWE-134) in Notepad++'s Find Results panel, triggered when the editor parses a malicious nativeLang.xml language pack file. It is not an AI/ML system vulnerability — Notepad++ is a general-purpose text editor occasionally present on developer and ML engineer workstations, but the flaw itself does not touch model serving, training pipelines, or any AI component. Exploitability is low: it requires local access, user interaction (opening/using search after installing a poisoned language pack), there is no EPSS score, no public exploit or Nuclei template, and it is not in CISA KEV. Action: track this as standard desktop software patch management rather than AI supply-chain risk — update Notepad++ past 8.9.3 per the vendor advisory, and only source language packs from the official Notepad++ channel, not community forums.

Sources: NVD CISA KEV notepad-plus-plus.org vulncheck.com

What is the risk?

Overall risk is low for most organizations. CVSS 4.4 (medium) reflects a local attack vector, low complexity, no privileges required, but mandatory user interaction and a confidentiality/availability-only impact (no code execution, no integrity impact). There is no evidence of active exploitation (not in CISA KEV, no SSVC decision recorded), no EPSS scoring available, and no public exploit code or scanner signature exists. The realistic worst case is an application crash (DoS) or leakage of stack/register contents that could aid a follow-on attack (e.g., ASLR bypass) — not direct data theft or remote compromise. This CVE carries negligible relevance to AI/ML system risk specifically; it should be triaged as routine endpoint software patching, not AI/ML supply-chain or model risk.

How does the attack unfold?

Malicious pack distribution

Attacker crafts a poisoned nativeLang.xml with format string payloads and distributes it via community channels posing as a legitimate language pack.

User installation

Victim installs the language pack into Notepad++ to change the editor's display language, unaware of the embedded payload.

Trigger via search

Victim performs a Find/search operation, causing the Find Results panel handler to interpret the format string.

Impact

The application crashes (denial of service) or leaks stack/register memory contents, potentially aiding further exploitation on the host.

Malicious pack distribution

Attacker crafts a poisoned nativeLang.xml with format string payloads and distributes it via community channels posing as a legitimate language pack.

User installation

Victim installs the language pack into Notepad++ to change the editor's display language, unaware of the embedded payload.

Trigger via search

Victim performs a Find/search operation, causing the Find Results panel handler to interpret the format string.

Impact

The application crashes (denial of service) or leaks stack/register memory contents, potentially aiding further exploitation on the host.

What systems are affected?

Package	Ecosystem	Vulnerable Range	Patched
Notepad++	—	—	No patch

Do you use Notepad++? You're affected.

How severe is it?

CVSS 3.1

4.4 / 10

EPSS

N/A

Exploitation Status

No known exploitation

Sophistication

Moderate

What is the attack surface?

AV Local

AC Low

PR None

UI Required

S Unchanged

C Low

I None

A Low

What should I do?

1 step

1) Upgrade Notepad++ to the version released after 8.9.3 per the vendor's official release notes (notepad-plus-plus.org/news/v894-released). 2) Only install language packs from the official Notepad++ distribution — treat community-sourced nativeLang.xml files as untrusted, and block their installation via endpoint policy if language packs aren't centrally managed. 3) For fleets managing developer workstations, verify installed Notepad++ versions via software inventory/EDR and flag any instance still on 8.9.3 or earlier. 4) Detection: monitor for unexpected Notepad++ crashes/access violations correlated with recent language pack changes; there is no known IOC or scanner signature published for this issue as of enrichment.

How is it classified?

DoS Data Leakage Supply Chain Plugin

Which compliance frameworks are affected?

Compliance analysis pending. Sign in for full compliance mapping when available.

Frequently Asked Questions

What is CVE-2026-6539?

This is a local format string injection (CWE-134) in Notepad++'s Find Results panel, triggered when the editor parses a malicious nativeLang.xml language pack file. It is not an AI/ML system vulnerability — Notepad++ is a general-purpose text editor occasionally present on developer and ML engineer workstations, but the flaw itself does not touch model serving, training pipelines, or any AI component. Exploitability is low: it requires local access, user interaction (opening/using search after installing a poisoned language pack), there is no EPSS score, no public exploit or Nuclei template, and it is not in CISA KEV. Action: track this as standard desktop software patch management rather than AI supply-chain risk — update Notepad++ past 8.9.3 per the vendor advisory, and only source language packs from the official Notepad++ channel, not community forums.

Is CVE-2026-6539 actively exploited?

No confirmed active exploitation of CVE-2026-6539 has been reported, but organizations should still patch proactively.

How to fix CVE-2026-6539?

1) Upgrade Notepad++ to the version released after 8.9.3 per the vendor's official release notes (notepad-plus-plus.org/news/v894-released). 2) Only install language packs from the official Notepad++ distribution — treat community-sourced nativeLang.xml files as untrusted, and block their installation via endpoint policy if language packs aren't centrally managed. 3) For fleets managing developer workstations, verify installed Notepad++ versions via software inventory/EDR and flag any instance still on 8.9.3 or earlier. 4) Detection: monitor for unexpected Notepad++ crashes/access violations correlated with recent language pack changes; there is no known IOC or scanner signature published for this issue as of enrichment.

What is the CVSS score for CVE-2026-6539?

CVE-2026-6539 has a CVSS v3.1 base score of 4.4 (MEDIUM).

What are the technical details?

Original Advisory

Notepad++ 8.9.3 contains a format string injection vulnerability in the Find Results panel handler that allows attackers to cause denial of service and information disclosure by crafting a malicious nativeLang.xml language pack file. Attackers can distribute a poisoned language pack through community channels that triggers format string interpretation when a user performs search operations, leading to access violations and potential leakage of stack or register contents.

Exploitation Scenario

An attacker crafts a malicious nativeLang.xml file embedding format string specifiers and distributes it through community plugin/language-pack sharing channels (forums, unofficial repositories), disguising it as a legitimate localization pack. A victim downloads and installs the file to change their Notepad++ UI language. The next time the victim performs a search operation that populates the Find Results panel, the format string is interpreted by the panel handler, causing an access violation — resulting in a crash (denial of service) or, in some conditions, disclosure of stack/register memory contents that could leak fragments of process memory useful for further exploitation on that host.

Weaknesses (CWE)

CWE-134 Use of Externally-Controlled Format String

CWE-134 — Use of Externally-Controlled Format String: The product uses a function that accepts a format string as an argument, but the format string originates from an external source.

[Requirements] Choose a language that is not subject to this flaw.
[Implementation] Ensure that all format string functions are passed a static string which cannot be controlled by the user, and that the proper number of arguments are always sent to that function as well. If at all possible, use functions that do not support the %n operator in format strings. [REF-116] [REF-117]

Source: MITRE CWE corpus.